Theme Article: Inclusive Data Experiences
An Age-based Study into Interactive Narrative
Visualization Engagement
Nina Errey
1*
, Yi Chen
2
, Yu Dong
3
, Quang Vinh Nguyen
4
, Xiaoru Yuan
5
, Tuck Wah Leong
6
, and Christy Jie Liang
1
1
School of Computer Science, University of Technology Sydney, Sydney, New South Wales, Australia
2
Beijing Key Laboratory of Big Data Technology for Food Safety, Beijing Technology and Business University, Beijing,
China
3
Advanced Interactive Technology and Application Laboratory, Computer Network Information Center, Chinese
Academy of Sciences, Beijing, China
4
Western Sydney University, Sydney, New South Wales, Australia
5
Peking University, Beijing, China
6
School of Computing Technologies, RMIT University, Melbourne, Australia
*
Nina Errey is the corresponding author.
Abstract—Research has shown that an audiences’ age impacts their engagement
in digital media. Interactive narrative visualization is an increasingly popular form of
digital media that combines data visualization and storytelling to convey important
information. However, audience age is often overlooked by interactive narrative
visualization authors. Using an established visualization engagement questionnaire,
we ran an empirical experiment where we compared end-user engagement to
audience age. We found a small difference in engagement scores where older age
cohorts were less engaged than the youngest age cohort. Our qualitative analysis
revealed that the terminology and overall understanding of interactive narrative
patterns integrated into narrative visualization was more apparent in the feedback
from younger age cohorts relative to the older age cohorts. We conclude this paper
with a series of recommendations for authors of interactive narrative visualization
on how to design inclusively for audiences according to their age.
D
emographic characteristics such as an audi-
ences’ age can impact how they interact and
engage with digital media [
1
,
2
]. Interactive
narrative visualization, or data storytelling is a form
of digital media that has been shown to compel and
explain information through an engaging end-user expe-
rience [
3
,
4
,
5
]. It has thus been used to communicate
critical topics such as the effects of climate change
or election outcomes. Despite conveying important
information, interactive narrative visualization is often
designed without audience age in mind. One potential
reason is that creating interactive narrative visualization
is a labor-intensive process that requires expertise
and creativity [
7
]. Researchers have therefore concen-
trated on aiding authors to alleviate the challenges
involved in interactive narrative visualization creation.
Consequently, researchers have yet to fully explore
the requirements of different audiences of interactive
XXXX-XXX © IEEE
Digital Object Identifier 10.1109/XXX.0000.0000000
narrative visualization.
In this study we attempt to address a fundamental
knowledge gap on the question of how audience age im-
pacts engagement in interactive narrative visualization.
We investigate if age significantly impacts engagement
in interactive narrative visualization. Moreover, we qual-
itatively analyze why differences exist. We aim to give
a concrete, evidence-based answer to whether authors
of narrative visualization should prioritize audience age
when designing interactive narrative visualization and
how this should be done.
To achieve our aim, we performed an empirical
experiment. We developed three narrative visualization
examples that employed different interactive narrative
patterns tailored specifically for engagement [
8
]. We
randomly assigned 2400 participants to one of the
three narrative visualization examples and measured
their engagement using VisEngage [
6
]. VisEngage is
a self-reporting questionnaire specifically developed to
measure engagement in visualization. The age groups
were split into generations. A younger audience with an
Published by the IEEE Computer Society
1
This article has been accepted for publication in IEEE Computer Graphics and Applications. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/MCG.2025.3591817
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: INRIA. Downloaded on December 11,2025 at 13:34:41 UTC from IEEE Xplore. Restrictions apply.
THEME ARTICLE
age of 18-27, a middle-younger audience between the
ages of 28-43, an older audience of 44-59, and finally,
the oldest age cohort consisting of over 60. These age
groups were determined by generation boundaries.
The results of our study revealed that there is a small
but significant difference between older audiences and
younger audiences’ engagement in narrative visualiza-
tion. From our qualitative analysis, it was found that
younger audiences were more observant of the interac-
tive techniques employed that encouraged engagement.
The terminology used by younger audiences was
distinctly different from their older counterparts, where
they described the cognitive processes involved in their
interactive engagement. Older audiences were not so
discerning and reported their negative engagement was
partially due to the interactive device causing confusion
or distraction. To our knowledge, this is the first study
that investigates age groups and interactive narrative
visualization engagement. This study contributes foun-
dational research on interactive narrative visualization
audiences. We conclude the work by presenting a series
of recommendations for designing interactive narrative
visualization inclusively based on our findings.
Related Work
Narrative Visualization
Narrative visualization is a popular subject in the
visualization community. Initial work on the topic was
pioneered by Segel and Heer, who coined the term
‘narrative visualization’[
7
]. They proposed a design
space outlining genres and structures [
7
]. Further foun-
dational research on narrative visualization established
an analytic framework by defining rhetorical techniques
and possible transitions for story-sequencing [
9
,
10
].
Recent advances in narrative visualization authoring
processes include examples such as generative AI co-
creation systems, visualization generation using natural
language queries, and machine-guided workflows.
While there is much research on the authoring
process of narrative visualization, its impact on au-
diences is relatively under-explored. When viewing
visualization it has been found that an audience’s
personal beliefs impact their viewing experience [
11
].
Furthermore, the story structure can influence end-
user engagement in a narrative visualization [
12
]. More
recently, a series of ‘narrative patterns’ were observed
in narrative visualization, where the author’s intent was
correlated with a narrative device integrated into a data-
driven story [
8
]. Most have been studied and shown,
with varying degrees of success, to encourage audience
engagement [3, 4].
No singular definition of interactive narrative visual-
ization has been established. In this study, we use a
broad, inclusive definition for interaction in visualization,
as described by Dimara and Perin [
13
]. Their definition
is thus, “Interaction for visualization is the interplay
between a person and a data interface involving a data-
related intent” [
13
]. To avoid being too vague, we refer to
interactive narrative visualization, where an interactive
modality such as scrolling, clicking, or inputting end-
user-generated data is a key element of the narrative
visualization. This differs from interactive narrative
visualization from narrative visualization, such as data
videos or data comics, where an interactive modality
is not required. In the next section, we investigate
engagement and how it is measured for visualization.
Engagement
The human-computer interaction (HCI) community has
long considered engagement a fundamental concept in
user-centered design. Visualization research, in compar-
ison, has relatively recently begun to seriously regard
engagement [
6
,
15
]. The definition of engagement is
often ambiguous and dependent on discipline. We adopt
an HCI definition of engagement, which centers on
the quality of the user experience and on the positive
aspects of the interaction, particularly the phenomena
associated with being captivated by technology. O’Brien
et al. listed dimensions of engagement as including aes-
thetic appeal, novelty, perceived challenge, feedback,
motivation, and affect [
14
]. Alternatively, engagement
can be viewed as a continuum from low to high [
15
].
Furthermore, engagement in narrative visualization has
been viewed in the context of flow and fluid interaction
[16].
Numerous methods have been proposed to mea-
sure engagement in the field of visualization. Boy et
al. evaluated engagement by analyzing time spent on
interaction and user input [
17
]. Nowak et al. used elici-
tation interviews to examine factors including emotional
affect and engagement in narrative visualization [
5
].
A purpose-built method for measuring engagement
in visualization was proposed by Hung and Parsons
named VisEngage [
6
]. VisEngage is a self-reporting
questionnaire based on the user-engagement scale
adapted for visualization [
6
]. Similar adaptations were
successful in other domains, such as social networking
applications and games. A questionnaire comprising 22
questions, VisEngage addresses 11 engagement char-
acteristics, where each characteristic corresponds to
two questions. VisEngage is a relatively robust method
to measure end-user engagement in visualization [6].
2
This article has been accepted for publication in IEEE Computer Graphics and Applications. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/MCG.2025.3591817
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: INRIA. Downloaded on December 11,2025 at 13:34:41 UTC from IEEE Xplore. Restrictions apply.
THEME ARTICLE
Age-based Research in Visualization
Previous work on different age cohorts in visualization
usually focused on either the very old or the very young.
For example, visualization research with children has
investigated pedagogical approaches for visual literacy
and visualization design. Visualization research into
elderly audiences has examined aspects of accessibility,
comprehension and perception. The lack of understand-
ing of the needs of different age cohorts is a known
and cited dilemma in visualization research [2].
In the HCI field, age-based research has found
significant differences between age groups. For ex-
ample, the time taken to perform input modalities of
end-users compared to their age group found that older
adults were significantly different from their younger
counterparts. Although the results were inconclusive,
strong evidence pointed to differences between older
and younger age cohorts when using devices such as
smart watches [18].
In a report by the Interactive Advertising Bureau in
the United Kingdom (UK) it was found that younger age
cohorts find interactive advertisements more appealing
than their older counterparts [
19
]. However, in a report
by the NN Group, unnecessary interactivity and flashy
graphics are found to be ‘annoying’ by young adults
[
20
]. The NN Group report explained that young adults
are digital natives, who are “people raised in a digital,
media-saturated world” and distinctly different to their
older counterparts. Young audiences deemed ‘digital
natives’ were more confident and less patient with user
interfaces, according to the NN Group Report [20].
Older adults are described as wary of technology.
For example, older adults are supposedly less likely
to prefer gamified user experiences and prefer text-
based content. Moreover, older adults are reported to be
less confident with user interfaces and are “hesitant to
explore” [
20
]. Research into usability for older audiences
is still lacking, where their preferences and behaviors
are not adequately considered [
2
,
20
]. While et al.
introduced the term GerontoVis, which encapsulates
data visualization design that primarily focuses on older
adults, which they describe as a largely overlooked
area of visualization research [
2
]. The contribution of
this work is to provide evidence-based guidelines for
inclusively designing interactive narrative visualization
targeted toward an age group which will ultimately result
in more effective interactive narrative visualization.
Research Method
We conducted a crowd-sourced study using three nar-
rative visualization examples as a stimulus to achieve
our research aim of investigating if and how, end-
user age impacts engagement in interactive narrative
visualization. To see the interactive narrative visual-
ization example code and raw data, please see the
supplementary material here.
Experiment Design
One of the primary challenges when designing an
evaluation experiment is the dichotomy of localization
and globalization. It is important to have results that
can be globalized and, therefore, universally applicable.
Conversely, it is necessary to have strict experiment
parameters to report concrete results. To address
this challenge, we developed three different narrative
visualization examples that were similar in length but
varied in topic. Each example is different in its data,
messaging, and interactive narrative pattern. The intent
of the integrated interactive narrative patterns was,
however, similar – engagement [9].
The research team iteratively developed three
narrative visualization examples. Each example was
inspired by publicly accessible interactive narrative
visualizations from reputable publishers. Publishers that
influenced our designs include The New York Times,
ABC Australia, and The Pudding. By no means the
only publishers of interactive narrative visualization,
each of the aforementioned publishers is commended
in online journalism awards for their interactive narrative
visualization.
Design One: ’Make a Guess’ The first narrative
visualization example, Design One, used a ‘Make a
Guess’ interactive narrative pattern [
9
]. This pattern
encourages engagement by stimulating the curiosity of
an audience [3].
The audience is asked to guess an answer to
a question, and the answer to the question is then
revealed, affirming or disaffirming the accuracy of their
answer. The objective of the ‘Make a Guess’ narrative
pattern is that the audience questions their perception
of reality by revealing a mismatch between perception
and the actual data. An example of the ‘Make a
Guess’ narrative pattern is a New York Times story
on education titled ‘You Draw It: How Family Income
Predicts Children’s College Chances.’
In our study, the narrative visualization design
example, which we refer to as Design One, ‘Make a
Guess’, was based upon a dataset from the WWF’s
Living Planet Report 2022. It opened by asking the
participant if they could answer this question; ‘What
do you think is the percentage of decline of wildlife
populations since 1970?’ Underneath the question was
a sliding bar set by default to 20% and a button stating,
3
This article has been accepted for publication in IEEE Computer Graphics and Applications. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/MCG.2025.3591817
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: INRIA. Downloaded on December 11,2025 at 13:34:41 UTC from IEEE Xplore. Restrictions apply.
THEME ARTICLE
FIGURE 1. A diagram of interactive narrative visualization design examples A) Design One: ‘Make a Guess’ B) Design Two:
‘Breaking the Fourth Wall’ and C) Design Three: ‘Exploration’
‘find out.’ Once the participant clicked on the button, if
the sliding bar was set to any number under 69%, the
participant would receive the same message - wildlife
decline was more than their estimate. If they estimated
above 69%, they were answered with a ‘you are close.’
The default sliding bar amount, set at 20%, was a
deliberate ploy for the user to estimate a lower value,
and thus be surprised by the correct answer. See Figure
1:A for a diagram of Design One.
Design Two: ’Breaking the Fourth Wall’ The second
narrative visualization example, Design Two, used the
‘Breaking the Fourth Wall’ interactive narrative pattern.
‘Breaking the Fourth Wall’ is a term often cited in
cinema and literature disciplines. In interactive narrative
visualization, a direct question is asked of the audience,
normally to input personal data. This creates a ‘self-
story connection,’ which has been found to encourage
engagement, as it includes the user within the story
[
4
]. The narrative visualization design example that
inspired this study was a finalist in the online journalism
awards. Published by the ABC Australia Story Lab, the
narrative visualization is titled ‘See how global warming
has changed the world since your childhood.’ In our
study, Design Two broke the fourth wall by asking the
user to input their name. Specifically, the user was
asked to ‘please enter your first name (this data will
not be stored)’ so that privacy concerns were availed
with the assurance that data relating to the user’s name
would not be stored. Design Two opened by stating,
‘How likely are you to be murdered by a serial killer?’
The user was asked to enter their first name click submit.
Once submitted, a screen appeared with ‘[name], have
you ever watched a horror movie and found yourself
scared? Should you be scared of serial killers?’ This
is followed by a basic bar chart explaining that ’1%
of homicides in the UK are about 5.’ The narrative
sequentially revealed itself as the user scrolled down
4
This article has been accepted for publication in IEEE Computer Graphics and Applications. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/MCG.2025.3591817
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: INRIA. Downloaded on December 11,2025 at 13:34:41 UTC from IEEE Xplore. Restrictions apply.
THEME ARTICLE
the screen.’ All data sources are referenced at the
end of the narrative visualization. See Figure 1:B for a
diagram of Design Two.
Design Three: ’Exploration’ We refer to the third
interactive narrative visualization design example as
‘Design Three.’ Differentiating from the previous two
examples, Design Three integrated a narrative pattern
that encouraged data exploration. The audience is
asked to freely explore data so that they can create
their narrative. Such an experience is described as
a ‘reader-driven’ narrative visualization. Design Three
was inspired by a narrative visualization that appeared
in a digital publication called ‘The Pudding.’ The specific
interactive narrative visualization was titled ‘A Visual
Guide to the Aztec Pantheon’, which explained Aztec
iconography. Similar in interface design to the Pudding
example, Design Three encouraged users to click
on the interface to explore information. The example
asked users, ‘Which personality type are you? The
Enneagram system is a theory that assigns everyone a
personality type.’ The user was then asked to click on
the personality types to find out which one they most
like. See Figure 1:C for a diagram of Design Three.
Survey Design
The survey instrument was adapted from the VisEngage
engagement questionnaire [
6
]. The survey instrument
contained 22 questions, where the 11 engagement
characteristics were allocated two questions each.
For clarity, the wording of each question mentions a
‘data story’ rather than a narrative visualization. The
participant could answer on a 7-point Likert scale
from strongly disagree to strongly agree. Examples
of the questions are as follows; ‘While using this data
story, I found its look and feel to be pleasing’ or ‘The
content or message of this data story was interesting
to me.’ As described by Hung and Parsons, an overall
engagement score can be achieved by adding together
the results of each question [
6
]. Strongly disagree is
allocated a one and strongly agree is allocated a 7.
Therefore, the maximum engagement score is 154,
corresponding to high engagement, and the minimum
is 22, corresponding to low engagement.
Experiment Procedure
We conducted the experiment on the Prolific crowd-
sourcing platform. The experiment was in three phases.
The first phase was where the participant exited Prolific
and moved to the Qualtrics survey platform. They
were asked to read and consent to the consent form,
where ethics details were attached. In the next step,
participants were asked for their Prolific ID, which was
automatically inserted, and an attention check question.
If the participant failed to consent, add their Prolific ID,
or failed the attention check, their token was revoked,
and they were returned to Prolific.
The second phase was where the participant was
asked to ‘please interact with the data story and
then answer the questions below.’ One of the three
randomly allocated interactive narrative visualization
designs was presented using an iFrame, where the
interactive narrative visualization was hosted on an
external server. An iFrame is an HTML element that
allows you to embed another HTML document. After
the iFrame, we posed a comprehension question to
ensure participants had interacted with the narrative
visualization. After the comprehension check question,
the participant answered the 22 engagement questions.
All questions were on a Likert scale, and all were
mandatory.
The final phase of the experiment was a qualitative
feedback question that asked, ‘Did you feel you were
engaged in the data story? Why or why not?’ This
question was not mandatory. The participant could then
either submit a response or move to the next step, which
returned them to the Prolific platform. See Figure 2 for
flow chart of the experiment procedure.
Participants
We split each age cohort according to what are often
described as ‘generations.’ Generational research is a
foundational topic in social sciences; however, we would
highlight that the labels used to describe generations
can be loaded with stereotypical connotations. The
objective of this research is not to perpetuate stereo-
types associated with generational labels, and therefore
we are not using the commonly used labels. We will
refer to each age cohort by their age and refrain from
using labels to diminish stereotypical connotations. The
youngest cohort consisted of ages ranging between
18-27. This age bracket saw participants born on or
after 1997. Due to limitations with the crowd-sourcing
platform, the youngest participant allowable age was 18.
The second cohort had ages ranging between 28-43,
where their birth year was on or between 1981 – 1996.
This age cohort came to adulthood during the first years
of the new millennium. Born between 1965- 1980, this
age cohort is between 44-59 years old. Finally, the
oldest cohort consisted of participants with ages 60+
with a birth year on or after 1965.
Our participant sample size was a result of a power
calculation with a goal of a 95% confidence level and
a 4% margin of error. The population was calculated
5
This article has been accepted for publication in IEEE Computer Graphics and Applications. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/MCG.2025.3591817
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: INRIA. Downloaded on December 11,2025 at 13:34:41 UTC from IEEE Xplore. Restrictions apply.
THEME ARTICLE
FIGURE 2. A flow diagram of experiment procedure. Experiment procedure: Prolific platform; ethics consent form; attention
check, interactive narrative visualization designs; comprehension check; engagement questionnaire and feedback question.)
TABLE 1. Combined engagement scores for all examples:
mean and standard deviation (SD) per age cohort.
Age Cohort Mean SD
18–27 112 15.8
28–43 110 15.5
44–59 109 17.1
60+ 109 16.8
based on the adult population size in the UK in 2022.
Our ideal sample size was calculated at approximately
601 participants per age cohort, therefore, with three
design examples with equally distributed participants
with 200 in each group, our total ideal sample size was
approximately 2400 participants.
Results
Hypothesis
We expected to observe differences between the four
age cohorts while factoring in the effect of the narrative
visualization design examples. We firstly affirmed if a
significant difference exists, specifically, our alternate
hypothesis was as follows:
H1: There is a significant difference in engagement
score and age cohort
Quantitative Analysis
Out of the 2400 participants, 77 failed the comprehen-
sion check. To investigate which age cohort had the
greatest engagement, our dependent variable needed
to be the overall engagement score. This was calculated
by adding together all question responses in the
VisEngage questionnaire as recommended by Hung
and Parsons [10].
We normalized the engagement score data remov-
ing extreme outliers. Outliers were identified with en-
gagement scores below 50, where we judged that their
extreme response patterns indicated likely response
bias. The number of outliers amounted to 43 partic-
ipants, less than 2% of participants. A Shapiro-Wilk
test was performed, and the result was not significant.
Approximately equal variances were tested using the
engagement score as the dependent variable in Lev-
ene’s test, the result of which was not significant. The
final number of participants totaled 2280 participants.
We ran a one-way ANOVA to compare the effect
of age and engagement score. The one-way ANOVA
revealed that there was a statistically significant dif-
ference in engagement between at least two groups
(F(3, 2426) = [2.81], p = [0.03]). Tukey’s HSD test
for multiple comparisons found that the mean value
of the engagement score was significantly different
between the 60+ age cohort and 18-24 age cohort
with a confidence interval of 95% (p = 0.05, 95% C.I.
= [-5.1, 0.02]). All other age cohorts between groups
means comparisons showed no significant differences.
Figure 3 shows a series of histograms of engagement
scores per age cohort for all combined design examples.
Figure 3 illustrates that the 18-24 age cohort has a
higher frequency of higher engagement scores relative
to the 60+ age cohort. Figure 4 compares engagement
scores of the 60+ age cohort and 18-24 age cohort,
by overlaying them on the same axis and highlighting
the difference. Table 1 shows the mean and standard
deviation per age cohort. It further illustrates the small
difference between the mean of each age cohort.
According to our test results we could accept our
research hypothesis (H1). We then measured effect
6
This article has been accepted for publication in IEEE Computer Graphics and Applications. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/MCG.2025.3591817
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: INRIA. Downloaded on December 11,2025 at 13:34:41 UTC from IEEE Xplore. Restrictions apply.
THEME ARTICLE
FIGURE 3. Series of histograms presenting each age cohort and frequency of engagement score.
FIGURE 4. Histogram comparing oldest and youngest age cohort and frequency of engagement score on the same axis.
size of age on engagement score. The effect size of
age cohort on engagement score, as measured by
Cohen’s f, was 0.06, indicating a small effect size
(95% C.I. = [0.01, 0]). We analyzed the interaction
effect of age cohort and narrative visualization design
example on participant engagement scores. The two-
way ANOVA revealed that there was no statistically
significant interaction between the effects of age cohort
and narrative visualization design example (F(6, 2355)
= [1.52], p = [0.17). Therefore, when factoring in the
narrative visualization design examples, they did not
interact with participant engagement score.
Finally, overall engagement in the interactive narra-
tive visualization was positive for all age cohorts. The
mean engagement score was 110 for all age cohorts
combined. Furthermore, the median engagement score
was 112. This data reveals that the majority of partici-
pants were positively engaged in interactive narrative
visualization.
Qualitative Analysis
We investigated the thought processes of participants
by analyzing their responses to a qualitative feedback
question. The survey instrument asked, ‘Did you feel
you were engaged in the data story? Why or why not?’
The aim was to shed light on the cognitive reasoning
and reflections of participant on their engagement with
the interactive narrative visualization. This question was
not mandatory, and we received 2278 responses. To
qualitatively analyze the data, we adopted an inductive
theming approach using latent theming. We inductively
coded lower-level themes determined by upper-level
themes. Initially we determined upper-level themes
by word frequency matched to potential engagement
related issues. Examples such as ‘design’ and ‘inter-
active’ were deemed as upper-level themes according
to their relative high frequency in the qualitative data.
Two coders independently coded responses, and any
inconsistencies were discussed.
7
This article has been accepted for publication in IEEE Computer Graphics and Applications. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/MCG.2025.3591817
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: INRIA. Downloaded on December 11,2025 at 13:34:41 UTC from IEEE Xplore. Restrictions apply.
THEME ARTICLE
Interactivity ‘Interactive’ was explicitly mentioned 7
times by the 60+ cohort. This was in contrast to the 18-
27 age cohort, who explicitly mentioned ‘interactive’ or
‘interactivity’ 49 times. We analyzed the exact phrases
that participants used in the 18-27 cohort relating to
interactivity. Our analysis revealed that the younger
audience attributed their engagement to interactivity,
for example, “yes because it was interactive” or I was
engaged as it was an interactive task” (both comments
from 18-27 age cohort, Design One). The interactive
device in Design Two, aimed to include the audience in
the story and thus encourage engagement. We coded
22 instances where the 18-24 age cohort recognized
that this was the aim of the interactive device in Design
2. For example, “Yes as it was interactive and by using
my name felt personal” (18-27 age cohort, Design Two).
In contrast, we coded 3 instances in the 60+ age cohort
specifically mentioning the interactive device in Design
Two. We found further evidence of perceptiveness in
Design Three from the 18-27 age cohort, where we
coded 18 responses that noticed the interactive device,
which in this case required clicking and exploring the
interface. For example, “I did feel I was engaged in
the data story as I had to click to find the information
as well as scroll for more information” (18-27 age
cohort, Design Three). The goal of the interactive device
in Design One, ‘Make a Guess’ was to illustrate a
mismatch between audience expectation and reality.
We coded 17 responses from the 18-27 age cohort
that referenced the interactive device. For example,
‘The decline is higher’ which made me feel engaged”
(18-27 age cohort, Design One). In contrast the 60+ age
cohort mentioned the interactive device in Design One
6 times. While we observed that 18-24 age cohort was
relatively more aware of the interactive devices this does
not mean other age groups were oblivious, only their
perceptive feedback was less frequent. For example, “I
felt engaged as there was an interactive question where
I could enter what I thought to be the answer. This made
the impact of learning the true answer heavier as I was
engaging with the story” (44-59 age cohort, Design One)
or “I found the personality types interesting flip over to
read and to associate the descriptions with the images
you gave for the personality” (60+ age cohort, Design
Three). When recognized by the older age cohorts,
there were 17 instances where the interactive device
had the opposite effect of encouraging engagement and
reported as a distraction or confusion, “I was distracted
by, initially, not realizing I had to scroll down the box to
gain more information” (60+ age cohort, Design Two)
or “I found the chart near the end distracting as it
seemed confusing.” (44-59 age cohort, Design Three).
In the 18-27 age cohort, there were 2 observable
reports of the interactive devices causing confusion
and none causing distraction. The two reports were
varied on the reasoning for why the interactive narrative
visualization was deemed confusing. The younger age
cohorts did report that they preferred an easy-to-use
interface, where the interactive device did not detract
from the user experience. We coded 57 responses
from the younger age cohorts that mentioned interface
functionality, in both positive and negative light. As
described here, “it would have been better if you did
not have to scroll down” (28-43 age cohort, Design
One), or it “felt like a bit of a gimmick, why not just
have the information under the pictures instead of
needing to click nine times” (18-27, Design Three).
The data suggests that regardless of the audiences’
age, usability influences engagement. For the older age
cohorts, however, the effects of poor usability cause
greater effect than a minor aggravation, but feelings of
confusion and distraction.
Cognition In older audiences, it has been shown that
complex visualizations can be cognitively demanding,
requiring users to remember and interpret multiple
information pieces simultaneously. We found evidence
that older audiences expressed a preference for less
complex visualization, for example, “I don’t find it par-
ticularly easy to interpret graphs or charts and I loathe
Venn diagrams, so, I possibly had to concentrate more
than other participants in order to ensure that I was
interpreting the information correctly” (60+ age cohort,
Design Two) or “I found the charts a little confusing
to begin with - possibly my age!” (44-59 age cohort,
Design One). Furthermore, older audiences mentioned
that they were required to revisit the narrative visualiza-
tion to fully comprehend it. We coded 9 instances where
participants from the older age cohorts mentioned they
missed data in the interactive visualization, this was
compared one instance that appeared in the 28-43
age cohort and none were observable in the 18-27
age cohort. “My biggest problem with it was the need
to scroll down. At first I didn’t realize there was more
data below” (60+ age cohort, Design One). This reveals
older audiences might miss crucial information if the
representation is too complex or the interactive device
is not clearly marked.
Aesthetic Appeal We examined the responses of
the participants who reported they were not engaged.
The primary reason, reported by the 18-27 age cohort,
was criticism of the aesthetic appeal of the narrative
visualizations. For example, Design One, had a black
background, which was described as ‘dated’, where for
example it was stated, “It felt quite outdated especially
8
This article has been accepted for publication in IEEE Computer Graphics and Applications. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/MCG.2025.3591817
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: INRIA. Downloaded on December 11,2025 at 13:34:41 UTC from IEEE Xplore. Restrictions apply.
THEME ARTICLE
with the colors” (18-27, Design One). Design Three,
was described as ‘cluttered’, for example, “No as the
text for the personality types 1-9 was cluttered” (18-27,
Design Three). We coded 12 instances in the 18-27 age
cohort commenting on the color palette of Design One.
Comparably, the 60+ age cohort mentioned the color
palette of Design One 3 times. These comments reveal
the importance of sound aesthetic design, particularly
for an interactive narrative visualization aimed at a
younger audience.
One difference observed between the older and
younger age cohorts was their preference toward text
integrated into the narrative visualization. Younger
audiences preferred less text that was divided into
smaller sections which they could control, for example
“I felt that the gradual reveal of information meant
that it was easier to compartmentalize statistics and
different pieces of information rather than looking at
a solid block of text, it felt more intuitive” (18-27,
Design One). We found 14 instances where the 18-
27 age cohort mentioned that they preferred text was
gradually revealed where the end-user could control the
pace. Older audiences asked for more information that
provided context to the interactive narrative visualization.
For example, “It was thought provoking and I felt it
needed more pages to explain what has been lost and
why” (60+ age cohort, Design One). 6 instances were
coded of the 60+ age cohort asking for more information.
The 18-27 age cohort had one observable instances
where they asked for more information.
Finally, we coded 1437 strongly positive responses.
This is in line with our quantitative data analysis, where
overall positive engagement was reported for all age
cohorts, however, in line with our quantitative analysis,
slightly less for the older age cohorts. The 60+ age
cohort reported 326 instances of positive engagement,
338 by the 44-59 age cohort, 382 by the 28-43 age
cohort, and 391 times by the 18-27 age cohort. This
reflected a generally positive opinion of engagement
in the narrative visualization examples across all age
cohorts.
Discussion and Future Work
The primary aim of this study was to find out if and
how audience age impacts engagement in interactive
narrative visualization. This research is a fundamental
step toward giving greater credence to the needs of
the audience when designing interactive narrative visu-
alization. We established that audience age did have
a small impact on end-user engagement in interactive
narrative visualization. The data revealed the greatest
difference was between the 60+ age cohort and the
youngest 18-27 age cohort.
In this section we present a series of recommenda-
tions for designing interactive narrative visualization for
specific age cohorts.
Designing Narrative Visualization for Older
Audiences
Older audiences are more attuned to usability difficulties.
We found that the 60+ age cohort reported a slightly
lower engagement score compared to the 18-27 age
cohort. We investigated the qualitative data to find
out why their engagement was lower. We found 17
instances where the 60+ age cohorts reported feeling
distracted and confused by the interactive narrative
patterns integrated into the narrative visualization.
These negative reactions could explain their lower
engagement scores of the older age cohort compared
to the youngest age cohort. It is important to note
that all age cohorts desired ease of use. The primary
difference between age cohorts was the extent of the
negative reaction to usability-related concerns. For
example, older age cohorts reported not recognizing
functionality in the interactive narrative visualization,
such as scrolling. We found 9 instances where it was
stated that data was missed due to usability difficulties
in the 60+ age cohort. Missing crucial data contained in
an interactive narrative visualization might result in the
central messaging being misconstrued. Younger age
cohorts recognized that they were required to scroll but
preferred that it was not required. The findings of this
study highlight the importance for narrative visualization
authors to prioritize usability for all audiences, however,
especially if the visualization is aimed at an older aged
audience. Moreover, for older age cohorts, important
information contained in an interactive narrative visual-
ization might be missed if the interactive device or data
representation is too complex.
Older audiences desire the option for more information.
It was found that older audiences asked for more
information and context to the narrative presented.
We suggest that authors of narrative visualization
provide the option for the end-user to access further
information about the presented topic. While it might not
be necessary for primary messaging contained within
the narrative visualization, more text provides context
to older audiences, where assumed knowledge might
not be present.
9
This article has been accepted for publication in IEEE Computer Graphics and Applications. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/MCG.2025.3591817
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: INRIA. Downloaded on December 11,2025 at 13:34:41 UTC from IEEE Xplore. Restrictions apply.
THEME ARTICLE
Designing Narrative Visualization for Younger
Audiences
Younger audiences understand interactive narrative
patterns. Qualitative data analysis indicated a stark
difference in the terminology used by the youngest age
cohort relative to the older age cohorts. Interactivity was
mentioned 49 times by the youngest age cohort, and
usually in a positive light. Furthermore, the youngest
age cohort seemed to have a perceptive understanding
of how the interactive narrative pattern achieved its
intent of encouraging engagement. For example, the
younger age cohort’s responses to Design Three,
‘Exploration’ specifically outlined how exploring the data
through clicking was more engaging than simply reading
it. Our study sheds light on the depth of understanding
of interactive devices held by the youngest age cohort,
illustrated by their descriptive feedback. The results of
this study do not disprove the NN Group report, which
stated it was a myth that young adults “crave multimedia
and innovative design.”[
20
] Rather, the results of this
study give credence to the fact that young adults
are accustomed and therefore more understanding of
interactive devices. Increased engagement is, therefore
largely due to the young adults’ ability to perceive
the intent of the author. Whereby recognizing that as
an audience, younger age cohorts are expected to
be engaged, therefore they are engaged. For future
authors of interactive narrative visualization targeted
towards younger age cohorts, it is recommended to
use interactive narrative patterns with explicit intent.
Duplicitous or superfluous use of interactive narrative
patterns would likely be recognized and thus could
result in lowered engagement.
Younger audiences appreciate aesthetics. Positive
aesthetic appreciation is a known contributor to en-
gagement [
14
]. Aesthetics were reported to directly
impact engagement for the younger age cohorts and
was attributed as their primary reason for not engaging
in the narrative visualization. Criticism regarding design
was overwhelming prevalent in this age cohort. These
criticisms included negative feedback on color, imagery
and layout. We observed that younger audiences
preferred an interface that was less cluttered and
thus easy to digest. It is recommended that authors
of narrative visualization segment their information
thoughtfully. Furthermore, while it is helpful to use
automated tools for narrative visualization generation,
the role of the narrative visualization author continues
to be of importance to aesthetically evaluate the overall
design and flow of the narrative visualization.
Design Recommendations for All Audiences
Interactive narrative visualization has a broad appeal.
One positive outcome of our study is the apparent
appeal of interactive narrative visualizations. The mean
engagement score for all age groups combined was 110.
The positive mean average indicates overall positive
engagement in interactive narrative visualization. In
addition, the qualitative analysis evidenced a largely
positive reaction, where 63% of feedback responses
reported a strongly positive engagement. This finding
evidences that as a communication medium, interactive
narrative visualization can engage a broad audience.
Authors of interactive narrative visualization should not
shy away from designing narrative visualization for older
audiences. The study presented here shows that, when
designed inclusively, interactive narrative visualization
is an engaging medium for all age demographics.
Future Research Opportunities
The sheer volume of data that was generated by the
large participant base in this study requires greater
inspection. Nuanced differences in the data were not
adequately considered as they were beyond the scope
of this work. For example, we have simply added
the engagement 22 questions, where each of the 11
engagement characteristics received two corresponding
questions, in the VisEngage questionnaire to achieve a
final engagement score. We added them together as it
was recommended by Hung and Parsons [
6
]. However,
it could be interesting to investigate if individual charac-
teristics appeal to age groups differently. Furthermore,
it stands to reason individual differences in design
examples impact the engagement of audiences. We
have provided the raw data from the experiment in the
supplementary material, where we encourage future
researchers to inspect and analyze the data in greater
detail. As the study of narrative visualization audiences
is a relatively emerging area of research, we suggest
other demographics that could be worthy as a focus of
investigation. Other demographics divided by education,
visual literacy, technological literacy or location could
prove to be interesting avenues of investigation. The
empirical evidence reported in this study was written
with the goal of making conclusions over a broad
audience divided merely by age cohort, however it is
clear age is just one of many audience characteristics
that can potentially impact engagement.
Ultimately this work results in more effective interac-
tive narrative visualization as it can better inform future
interactive narrative visualization design and research.
Empirical visualization research can overlook the age
of their participant base, this study shows that age
10
This article has been accepted for publication in IEEE Computer Graphics and Applications. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/MCG.2025.3591817
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: INRIA. Downloaded on December 11,2025 at 13:34:41 UTC from IEEE Xplore. Restrictions apply.
THEME ARTICLE
can impact experiment results and should be reported
[
2
]. We hope the findings of this study encourages
authors and researchers to seriously consider their
audience when designing or researching interactive
narrative visualization in the future, where age is but
one of many audience characteristics that should be
considered.
Limitations
There are several limitations to this study. The foremost
limitation is that there are but three narrative visualiza-
tion design examples. Optimally, we would have used a
multitude of examples. However, the scope of this study
dictated a limit of three. It should be noted, however, that
each design example uses one of the three interactive
narrative patterns that are described for engagement [
9
].
The objective of this study was a broad approach, where
we used differing topics, narrative patterns, and designs.
It is unfeasible to study all possible combinations of
topics, narrative patterns, and designs. We believe that
the three examples we developed were adequate to
achieve our study’s aims.
Another notable limitation is that this study is only
representative of an audience based in the UK. The
availability of the oldest age cohort from countries
outside the UK and the US was specifically challenging
and disappointing to the international research team.
The uneven distribution of older participant country
locations resulted in a decision to focus the experiment
on participants from the UK. Rather than a skewed
result, we prefer our results to concretely representing
the behaviors of peoples from one locale. Furthermore,
the premise of our study is to question whether different
demographics engage differently with narrative visual-
ization; therefore, it stands to reason that the locale
of participants might impact study results. Comparing
audience engagement across multiple countries is
outside the scope of this work.
Our study’s participant pool was recruited from an
online crowd-sourcing platform. Recruiting participants
from an online platform necessitated a level of technical
proficiency from participants. Furthermore, the study
required participants undertake the experiment on
a desktop computer. These factors resulted in our
study’s participant pool being skewed toward more
technically proficient participants. Future work could
consider recruiting participants from offline sources
where a lack of technical proficiency might impact the
study outcomes.
For future researchers, we have provided our de-
signs and code from the interactive narrative visual-
ization examples in the supplementary material. We
encourage researchers to replicate this study using
varied design examples, alternative languages or using
alternative recruitment strategies.
Conclusion
To communicate effectively, content authors are re-
quired to recognize the needs, preferences, and be-
haviors of their intended audience. The outcomes of
this study suggest that audience age impacts their
engagement in interactive narrative visualization. Older
audiences that are in the 60+ age cohort find that
interactive narrative patterns integrated into narrative
visualization cause usability difficulties. Younger age
cohorts do not experience the same response when pre-
sented with interactive narrative patterns. Younger age
cohorts recognize and appreciate interactive narrative
patterns and are thus more engaged than their older
counterparts. Our results lead to valuable implications
for designing future interactive narrative visualization,
where we encourage authors to give greater consid-
eration to their audience when designing interactive
narrative visualization.
REFERENCES
1.
J. Johnson and K. Finn, Designing user interfaces
for an aging population: Towards universal design.
Morgan Kaufmann, 2017.
2.
Z. While, R. J. Crouser, and A. Sarvghad, “Geron-
tovis: Data visualization at the confluence of aging,”
in Computer Graphics Forum, vol. 43, no. 3, pp.
e15101, 2024.
3.
M. Rogha, S. Sah, A. Karduni, D. Markant, and
W. Dou, “The impact of elicitation and contrast-
ing narratives on engagement, recall and attitude
change with news articles containing data visual-
ization,” IEEE Trans. Vis. Comput. Graph., vol. 30,
no. 7, pp. 4375–4389, 2024.
4.
Y. Shi, T. Gao, X. Jiao, and N. Cao, “Breaking
the fourth wall of data stories through interaction,”
IEEE Trans. Vis. Comput. Graph., vol. 29, no. 1,
pp. 972–982, 2022.
5.
S. Nowak, L. Bartram, and T. Schiphorst, “A micro-
phenomenological lens for evaluating narrative
visualization,” in Proc. BELIV, pp. 11–18, 2018.
6.
Y.-H. Hung and P. Parsons, “Assessing user en-
gagement in information visualization,” in CHI EA,
pp. 1708–1717, 2017.
7.
E. Segel and J. Heer, “Narrative visualization:
Telling stories with data,” IEEE Trans. Vis. Comput.
Graph., vol. 16, no. 6, pp. 1139–1148, 2010.
8.
B. Bach et al., “Narrative design patterns for data-
11
This article has been accepted for publication in IEEE Computer Graphics and Applications. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/MCG.2025.3591817
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: INRIA. Downloaded on December 11,2025 at 13:34:41 UTC from IEEE Xplore. Restrictions apply.
THEME ARTICLE
driven storytelling,” in Data-driven storytelling, AK
Peters/CRC Press, 2018, pp. 107–133.
9.
J. Hullman et al., “A deeper understanding of
sequence in narrative visualization,” IEEE Trans.
Vis. Comput. Graph., vol. 19, no. 12, pp. 2406–
2415, 2013.
10.
J. Hullman and N. Diakopoulos, “Visualization
rhetoric: Framing effects in narrative visualization,”
IEEE Trans. Vis. Comput. Graph., vol. 17, no. 12,
pp. 2231–2240, 2011.
11.
E. M. Peck, S. E. Ayuso, and O. El-Etr, “Data
is personal: Attitudes and perceptions of data
visualization in rural Pennsylvania,” in Proc. CHI,
pp. 1–12, 2019.
12.
J. Liem, C. Perin, and J. Wood, “Structure and
empathy in visual data storytelling: Evaluating their
influence on attitude,” in Computer Graphics Forum,
vol. 39, no. 3, pp. 277–289, 2020.
13.
E. Dimara and C. Perin, “What is interaction for data
visualization?,” IEEE Trans. Vis. Comput. Graph.,
vol. 26, no. 1, pp. 119–129, 2019.
14.
H. L. O’Brien et al., “Developing and evaluating
a reliable measure of user engagement,” Proc.
ASIS&T, vol. 45, no. 1, pp. 1–10, 2008.
15.
N. Mahyar, S.-H. Kim, and B. C. Kwon, “Towards
a taxonomy for evaluating user engagement in
information visualization,” in Workshop on Personal
Visualization, vol. 3, no. 2, p. 4, 2015.
16.
S. McKenna et al., “Visual narrative flow: Exploring
factors shaping data visualization story reading
experiences,” in Computer Graphics Forum, vol. 36,
no. 3, pp. 377–387, 2017.
17.
J. Boy, F. Detienne, and J.-D. Fekete, “Storytelling
in information visualizations: Does it engage users
to explore data?,” in Proc. CHI, pp. 1449–1458,
2015.
18.
Z. While and A. Sarvghad, “Dark mode or light
mode? Exploring the impact of contrast polarity on
visualization performance between age groups,” in
IEEE VIS, pp. 211–215, 2024.
19.
IAB.UK, “Interactive ads more appealing to
teens, study shows,” 2024. [Online]. Available:
https://www.iabuk.com/news-article/interactive-
ads-more-appealing-teens-study-shows
20.
Nielsen Norman Group, “Designing for young
adults (ages 18–25),” 2017. [Online]. Avail-
able: https://www.nngroup.com/reports/designing-
for-young-adults/
12
This article has been accepted for publication in IEEE Computer Graphics and Applications. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/MCG.2025.3591817
© 2025 IEEE. All rights reserved, including rights for text and data mining and training of artificial intelligence and similar technologies. Personal use is permitted,
but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: INRIA. Downloaded on December 11,2025 at 13:34:41 UTC from IEEE Xplore. Restrictions apply.